Enable pileup modelling in MLPF

[jpata]

For the CMS model, we have a per-particle is_pu target that is varying between 0 for particles from the hard event and 1 for particles from PU, and inbetween for merged particles.

It looks like this target could be used to make MLPF reconstruct a per-particle PU flag, or potentially restrict / weight the loss function to the hard event.